Pedestrian Density Analyzer – Project Story

🚀 Inspiration

After talking with Diego and Marcos, who were planning to build a pedestrian sensor, I wanted to create a website frontend to complement their hardware project. My goal was to implement web development and AI tools that would allow an admin to gain insights from multiple street intersections in real time.

📚 What I Learned

I learned a lot about the limitations of web development frameworks when integrating different services. At first, I attempted to use Cloudflare AI Workers, MongoDB, and Next.js together. This proved to be far more challenging than expected, but it gave me valuable insights into:

• How websites are structured and organized
  
• The intricacies of website hosting and deployment
  
• Where certain frameworks excel and where they struggle in real-world integration
  

🛠️ How I Built My Project

After several iterations and failed attempts, I discovered Chef by Convex, which allowed me to generate the website more effectively. This gave me a working foundation where I could integrate the pedestrian data entry, AI analysis, and visualization features into a functional dashboard.

⚡ Challenges I Faced

Some of the biggest challenges included:

• Cloudflare AI Workers limitations: Since Workers run in a V8 isolate instead of a traditional Node environment, it was very difficult to connect API calls between the Workers and MongoDB.
  
• Integration complexity: While I was able to connect Next.js to MongoDB, it wasn’t clear how to merge this with the Worker setup. For example:
  
  • Sending prompts via Next.js often wasn’t registered
    
  • Receiving responses from the Worker, however, worked fine
    
• Balancing different frameworks that were not designed to work seamlessly together, which slowed development but also gave me a much deeper understanding of system architecture.
  

Built With

• convex
• react
• vite